Pinned Repositories
driver-secondary-action-recognition
Driving-Environment-Detection
Driving environment detection using a multiclass classifier
Event-Correlation
Inferencing for external driving events such as a lane change based on gaze estimation and object detection input using a dash cam and a face camera.
gaze-fixation-and-object-saliency
This repository is related to estimating the driver's attention to the outside scene view as a point of gaze (PoG) w.r.t the gaze angles extracted from the driver facing view. We also explore analysis of gaze saliecny in form of heatmaps and yarbus plots.
GNN-based-intersection-safety
head-pose-detection
Intersection-Detection
Intersection vs Non-Intersection Calssfication based on Images/Videos
object-detection
This package contains configuration files and trained models using the MMdetection repo. The trained models can detect objects of transportation interest.
occupant_detection_classification
This repository provides code to detect people in car-cabin images and classify the type of passengers in each image (driver, front-seat passenger, back-seat passenger).
Segmentation-and-detection-of-work-zone-scenes
This project is concerned with the automatic detection and analysis of work zones (construction zones) in naturalistic roadway images. An underlying motivation is to identify locations that may pose challenges to advanced driver-assistance systems or autonomous vehicle navigation systems. We first present an in-depth characterization of work zone scenes from a custom dataset collected from more than a million miles of naturalistic driving data. Then we describe two ML algorithms based on the ResNet and U-Net architectures. The first approach works in an image classification framework that classifies an image as a work zone scene or non-work zone scene. The second algorithm was developed to identify individual components representing evidence of a work zone (signs, barriers, machines, etc.). These systems achieved an F{0.5} score of 0.951 for the classification task and an F1 score of 0.611 for the segmentation task. We further demonstrate the viability of our proposed models through salience map analysis and ablation studies. To our knowledge, this is the first study to consider the detection of work zones in large-scale naturalistic data. The systems demonstrate potential for real-time detection of construction zones using forward-looking cameras mounted on automobiles.
Virginia Tech Transportation Institute's Repositories
VTTI/gaze-fixation-and-object-saliency
This repository is related to estimating the driver's attention to the outside scene view as a point of gaze (PoG) w.r.t the gaze angles extracted from the driver facing view. We also explore analysis of gaze saliecny in form of heatmaps and yarbus plots.
VTTI/occupant_detection_classification
This repository provides code to detect people in car-cabin images and classify the type of passengers in each image (driver, front-seat passenger, back-seat passenger).
VTTI/Intersection-Detection
Intersection vs Non-Intersection Calssfication based on Images/Videos
VTTI/Event-Correlation
Inferencing for external driving events such as a lane change based on gaze estimation and object detection input using a dash cam and a face camera.
VTTI/head-pose-detection
VTTI/object-detection
This package contains configuration files and trained models using the MMdetection repo. The trained models can detect objects of transportation interest.
VTTI/Segmentation-and-detection-of-work-zone-scenes
This project is concerned with the automatic detection and analysis of work zones (construction zones) in naturalistic roadway images. An underlying motivation is to identify locations that may pose challenges to advanced driver-assistance systems or autonomous vehicle navigation systems. We first present an in-depth characterization of work zone scenes from a custom dataset collected from more than a million miles of naturalistic driving data. Then we describe two ML algorithms based on the ResNet and U-Net architectures. The first approach works in an image classification framework that classifies an image as a work zone scene or non-work zone scene. The second algorithm was developed to identify individual components representing evidence of a work zone (signs, barriers, machines, etc.). These systems achieved an F{0.5} score of 0.951 for the classification task and an F1 score of 0.611 for the segmentation task. We further demonstrate the viability of our proposed models through salience map analysis and ablation studies. To our knowledge, this is the first study to consider the detection of work zones in large-scale naturalistic data. The systems demonstrate potential for real-time detection of construction zones using forward-looking cameras mounted on automobiles.
VTTI/driver-secondary-action-recognition
VTTI/Driving-Environment-Detection
Driving environment detection using a multiclass classifier
VTTI/GNN-based-intersection-safety